Activation of corticospinal neurons by transcranial magnetic stimulation (TMS) is complex, the most notable features being recurrent discharges attributed to indirect activation (I-waves) at intervals of about 1.6 ms, which may vary dependent on coil type and orientation (e.g. Di Lazzaro et al. 2001). The interval between the I-waves is interesting since stimulation of the median nerve evokes synchronous oscillatory discharges in the somatosensory cortex at about 600 Hz (Gobbele et al. 1998), thus the interval between successive waves is about 1.6 ms. It is unknown whether these phenomena are linked. We have investigated this in the current experiments.

Experiments were performed on sixteen healthy volunteers with informed consent (11 males, 5 females) aged 20-43 years. The University of Cambridge Human Biology Research Ethics Committee approved the experiments. In each subject we first recorded the EEG response to median nerve stimulation (1.0 X motor threshold at the wrist). The oscillatory potentials at ~600 Hz in the somatosensory cortex were identified in the averaged (> 2000 sweeps) EEG response after digital filtering (500-700 Hz bandpass). After identifying the timing of the major peaks, an experimental protocol was set up so that TMS stimuli could be delivered at one of eight intervals spaced at 0.4 ms intervals after the median nerve stimulus, to span the first two complete cycles of the EEG oscillation. Stimuli at the different intervals were delivered in random sequence, interspersed with TMS only stimuli. Muscle-evoked potentials (MEPs) were recorded from hand muscles (thenar and first dorsal interosseus) and forearm muscles (extensor digitorum and flexor digitorum superficialis).

For initial analysis we examined whether the amplitudes of MEPs recorded at the eight points showed two peaks or two troughs, separated by 1.6 ms. 49 % of the data sets fulfilled this criterion, whereas by binomial distribution 21 % of data sets would be expected to. A Monte Carlo test showed that the presence of a modulation was highly probable (P < 0.01). Grouped analysis was undertaken by combining data across subjects aligned to the peak of the EEG burst in each case. The data for the thenar muscles showed a clear tendency for modulation at ~600 Hz, which was shown to be highly significant (P < 0.01) by a Monte Carlo test and Fourier analysis of the resultant waveforms.

The results suggest that the oscillatory activity set up in somatosensory cortex by median nerve stimulation is correlated with excitability changes in motor cortex, seen as modulations of the amplitude of MEPs evoked by TMS at different intervals after the median nerve stimulus. These findings may have implications for the origin of I-waves and sensorimotor integration.